The present invention relates to a method for, tuning a HF source, which oscillates at a basic frequency, to a narrow-band component, in particular for tuning a microwave source of an electron spin resonance spectrometer to a resonator, comprising the procedural steps of
tuning periodically the frequency of the HF signal supplied to the narrow-band component over a predetermined tuning range; PA1 detecting a first signal corresponding to the power of the HF signal absorbed by the narrow-band component; PA1 plotting the first signal as a function of the frequency within the tuning range in a manner such that the narrow-band power absorption range becomes visible as a mark in the representation of the first signal; and PA1 resetting the basic frequency so that the mark comes to lie substantially at the center of the tuning range. PA1 first means for tuning periodically the frequency of the HF signal supplied to the narrow-band component over a predetermined tuning range; PA1 second means for detecting a first signal corresponding to the power of the HF signal absorbed by the narrow-band component; PA1 third means for plotting the first signal as a function of the frequency within the tuning range in a manner such that the narrow-band power absorption range becomes visible as a mark in the representation of the first signal; and PA1 fourth means for resetting the basic frequency so that the mark comes to lie substantially at the center of the tuning range.
The present invention further relates to a device for tuning a HF source, which oscillates at a basic frequency, to a narrow-band component, in particular for tuning a microwave source of an electron spin resonance spectrometer to a resonator, having
A method and a device of the before-described kind have been known.
Electron spin resonance spectrometers of the conventional type usually use a reflex klystron as a microwave source. As is generally known, a reflex klystron is a component which enables the frequency of the output signal to be varied in two ways. On the one hand, coarse tuning of the frequency is possible by varying the volume of the resonator in the klystron, for example by means of a screw which is accessible from the outside. On the other hand, however, fine tuning of the klystron is possible by varying the reflector voltage.
If one now regards, by way of example, a reflex klystron oscillating in the X-band, it is possible in this case to obtain a tuning range of several 100 MHz with the usual mechanical tuning procedure, whereas electric fine tuning through the reflector voltage is limited to a tuning range of a few 10 MHz.
In the case of the before-mentioned application of electron spin resonance spectrometers, the microwave source feeds a resonator, which usually is designed as a cavity resonator. Cavity resonators have a very high unloaded Q in the unloaded condition so that tuning of the resonator to the microwave source, or vice versa, does not present any particular problems.
Now, it is a particularity of electron spin resonance spectrometers that their cavity resonators are loaded with samples. As long as the samples in question are solid samples having only low dielectric losses, the dielectric properties of the cavity resonator are not influenced significantly, which means that neither its oscillating frequency nor its Q value will vary substantially. Tuning is, therefore, connected with little problems in these cases.
Greater tuning problems are encountered, however, when instead of solid samples liquid samples with correspondingly high dielectric losses are to be measured, since introduction of such dissipative samples leads to quite significant variations in both the resonance frequency of the resonator and its Q value.
In order to enable the microwave source and the resonator to be mutually tuned also in these cases it is necessary to make full use of the mechanically tunable range of reflex klystrons, for example.
In the case of conventional electron spin resonance spectrometers, one proceeds for this purpose as follows:
Once the sample has been introduced into the resonator, the klystron is switched on, and this initially in an operating mode which only serves for mutual tuning of the klystron and the resonator. In this operating mode, the reflector voltage of the klystron is periodically tuned by means of a saw-tooth voltage. At the same time, the power absorption in the resonator is measured by tapping a corresponding diode current at the microwave detector--containing the resonator--of the microwave bridge that is anyway provided.
The saw-tooth voltage serves simultaneously for setting the horizontal deflection of a visual display unit, typically a cathode-ray tube. The vertical deflection is then supplied with a signal corresponding to the diode current. This has the result that a so-called "mode curve" appears on the visual display unit, i.e. a bell-shaped curve representing the power output of the klystron across the tuning range that can be covered by varying the reflector voltage. Now, when the resonance frequency of the resonator loaded with the sample is within the tuning range, this is indicated by an incision in the bell-shaped mode curve. The incision may be inside or outside of the maximum of the mode curve, which means that the resonator is not supplied with the maximally possible microwave power. On the other hand, however, the shape of the incision, in particular its width and its depth, provide a measure for the resonator Q and for the Q value of the resonator coupling circuit.
By mechanically adjusting the klystron it is now possible to displace the entire mode curve until the incision comes to lie at the center of the mode curve and, thus, in the maximum of the possible power output of the klystron. The same procedure is possible by analogy if, when switching on the spectrometer, the resonance of the resonator loaded with a sample is completely outside of the electric tuning range, as indicated by the mode curve on the visual display unit. In this case, one can initially proceed to mechanically detune the klystron with a view to finding the resonance point of the resonator which will then, during mechanical tuning of the klystron, move from the one or from the other side of the mode curve into the displayed image where it can then be centered in the manner described above.
The above procedure can of course also be automated; the above description has been given only for the sake of greater clarity.
As microwave component technology progressed, one has made numerous attempts at replacing the conventional microwave tubes, i.e. in particular the before-mentioned klystron, by semiconductor oscillators, in particular by Gunn oscillators. This measure was intended to eliminate the relatively complex circuitry of microwave tubes, i.e. the extremely high reflector voltages, the tube heating, and the like, with all the relevant thermic problems.
However, microwave semiconductor oscillators are not suited without qualifications for replacing microwave tubes in electron spin resonance spectrometers.
With Gunn oscillators, for example, it is not possible to tune them electrically over a wider frequency range of several 10 MHz, the tuning range of such oscillators being limited today to clearly less than 10 MHz. However, this relatively small tuning range is insufficient in those cases of electron spin resonance measurements where the resonator Q has been diminished so considerably that the effective bandwidth of the resonator has moved to a range of 10 MHz or above. It is then of course no longer possible to obtain a conventional mode curve, the electric tuning range of a Gunn oscillator being insufficient to cover the relatively wide bandwidth of the resonator loaded with the sample.
One has of course also considered in this connection to make use of the possibility, which in fact exists, of tuning Gunn oscillators mechanically, by varying the length of the resonator of the Gunn oscillator. While in principle such variation can be achieved by means of magnetostrictive or electrostrictive elements, or by electrodynamically excited diaphragm arrangements, trials in this direction have been little promising in practice, and this in particular due to the associated generation of noise.